1. Tatum, J., Apuleius and ‘The Golden Ass’ (Ithaca NY and London 1979), 19Google Scholar.

2. Anderson, G., Eros Sophistes: Ancient Novelists at Play (Chico CA 1982), 75Google Scholar.

3. Sandy, G.N., ‘Book 11: Ballast or Anchor’ in Hijmans, B.L. and van der Paardt, R.T. (eds), Aspects of Apuleius’ Golden Ass (Groningen 1978), 123–40Google Scholar, at 123.

4. Hooper, R.W., ‘Structural Unity in the Golden Ass’, Latomus 44 (1985), 398–401Google Scholar, at 398.

5. Schlam, C., ‘Platonica in the Metamorphoses of Apuleius’, TAPA 101 (1970), 477–87Google Scholar; Heller, S., ‘Apuleius, Platonic Dualism, and Eleven’, AJP 104 (1983), 321–39Google Scholar.

6. Schlam, , ‘The Curiosity of the Golden Ass’, CJ 64 (1968), 120–125Google Scholar; Sandy, , ‘Knowledge and Curiosity in Apuleius’ Metamorphoses’, Latomus 31 (1972), 179–83Google Scholar; Tatum, , ‘Apuleius and Metamorphosis’, AJP 93 (1972), 306–13Google Scholar.

7. Smith, W.S., ‘The Narrative Voice in Apuleius’ Metamorphoses’, TAPA 103 (1972), 513–34Google Scholar; Heath, J.R., ‘Narration and Nutrition in Apuleius’ Metamorphoses’, Ramus 11 (1982), 57–77CrossRefGoogle Scholar; James, P., Unity in Diversity: A Study of Apuleius’ Metamorphoses (Hildesheim 1987)Google Scholar. But the outstanding work in this category — indeed the most absorbing, instructive and stimulating work I have ever read on Apuleius — is Winkler, J.J., Auctor & Actor: A Narratological Reading of Apuleius’s The Golden Ass (Berkeley and Los Angeles 1985)Google Scholar. What follows will show the extent to which I am indebted to Winkler’s insights.

8. Division of a work into eleven books is otherwise unknown until Rufinus (late fourth century); see R. Heine, ‘Picaresque Novel Versus Allegory’ in Hijmans and van der Paardt (n.3 above), 25–42, at 37. Lavagnini, B., Studi sul romanzo greco (Messina and Florence 1950), 138Google Scholar, suggests that the first ten books correspond to the ten days of preparatory fasting (cf. Met. 11.23 and 28) and the eleventh to the day of initiation; this works for Book 11 perhaps, but hardly for the first ten, in which self-indulgence is far more prominent than abstinence. More attractive is Heller (n.5 above, 336), whose thesis is based on the Pythagorean significance of the number 10: ‘The first ten books represent the world or cosmos, while the eleventh book represents the transcendent god, over and above the world, to whom Isis leads the elect.’

9. Winkler (n.7 above), 207.

10. For a useful summary and criticism of these attempts see Winkler (n.7 above), 228–47.

11. See e.g. Winkler (n.7 above), 126: ‘Insofar as The Golden Ass makes us ask hard questions and does not supply authoritative answers, it may be called Platohic/Socratic … [It sets up] a Platonic dialogue between author and reader rather than among characters in the script.’ Winkler goes on to describe the work as ‘a shrewd kind of trick’ (215); the ending is one of ‘unresolved ambiguity’ (227) and the eleventh book as a whole manifests a ‘well-contrived balance of indeterminacy’ (245). Cf. J. Amat, , ‘Sur quelques aspects de l’esthétique baroque dans les Métamorphoses d’Apulée’, REA 74 (1972), 107–52CrossRefGoogle Scholar, at 151: ‘En ce sens le livre onze n’est pas un achèvement total, mais une solution possible’; Heath (n.7 above), 73: ‘The reader goes away empty-handed, not sure where he has been led or why. What was the purpose of the journey? He has been given something — a puzzle — in place of what had been expected. How are we to understand this reaction, this impression that we have in fact misinterpreted the literary intentions and narrative subtleties of Apuleius’ novel? Or could the answer simply be to have asked the question?’

12. Eur. And. 1273ff.

13. It is typical of this narrating ‘I’ that he presents the viewpoint appropriate to the time when the events took place, leaving the reader to make the necessary adjustments in the light of future experience. So the ‘lived happily ever after’ atmosphere of Charite’s and Tlepolemus’ wedding night (7.13f.) contains no hint of the violent end that will shortly befall this marriage (8.1–14); and the clear indication that caelestis prouidentia (‘heavenly providence’) has intervened at 9.27 to enable Lucius to get his own back on the miller’s wife by exposing her hidden lover and giving the miller grounds for divorce gives an added element of shock to the fact that it is the wicked wife who wins in the end (9.29–31). Cf. Winkler’s discussion of ‘before-and-after attitudes’ (n.7 above, 143–45).

14. Cf. Tatum (n.l above, 42) on the significance of the Risus Festival: ‘It serves as a kind of play-within-a-play, dramatising the true nature of the comedy of The Golden Ass … [and it] holds up a mirror to the readers of the novel; in the Schadenfreude of Lucius’ tormentors we may see a reflection of ourselves.’ But, as I shall argue, if all we are experiencing is titillation at the humiliation of others, then we are misreading.

15. On the distinction between first- and second-time readers see e.g. Winkler (n.7 above), 9–14.

16. The alert reader will also have his/her suspicions aroused by the legal impropriety (not to say the impossibility) of a free Roman citizen of Lucius’ exalted social status (on which see Mason, H.J., ‘The Distinction of Lucius in Apuleius’ Metamorphoses’, Phoenix 37 [1983], 135–43CrossRefGoogle Scholar) being threatened with torture and crucifixion (3.9) by a hastily convened court in a provincial town. See Millar, F., ‘The World of the Golden Ass’, JRS 72 (1981), 63–75Google Scholar, esp. 70f.; although after discussing this question Millar does add the salutary warning that ‘nothing is more illusory than the idea that the real world of the Roman Empire presents us with a clearly defined and intelligible system of public law and administration’ (71).

17. That second reading is demanded by this text is shown in the fact that it is only after this review that the humour in the circumstantial embellishment of Lucius’ defence (3.4–6) with its ludicrously inappropriate Socratic allusion (non tarn impunem me uerum etiam laudabilem pubike credebam fore, ‘I thought I would not only go unpunished but would actually be deemed worthy of public praise’ — cf. Plato Apol. 36d) can be fully appreciated. On the parodic aspects of this speech and the trial generally cf. Walsh, P.G., The Roman Novel (Cambridge 1970)Google Scholar, 58f.; van der Paardt, R.T., L Apuleius Madaurensis, The Metamorphoses: A Commentary on Book III (Amsterdam 1971)Google Scholar, 64f, notes that the speech is structured according to the norms of forensic rhetoric but has little to say about the humour thereby generated.

18. Cf. Thomas, J., Le dépassement du quotidien dans l’Énéide, les Métamorphoses d’ Apulée et le Satiricon (Paris 1986)Google Scholar, 39f, who interprets this episode as making the general point that in the phenomenal world appearances are deceptive: ‘Le ton est donné: le monde qui nous entoure est trompeur, le sens profond de notre vie et de ses aventures nous échappe, si nous nous en tenons aux apparences.’

19. Stephenson, W.G., ‘The Comedy of Evil in Apuleius’, Arion 3.3 (1964), 87–93Google Scholar, sees in the Festival of Laughter a foreshadowing of what he regards as the main theme of the novel, viz. that only divine intervention can subdue the cruelty and evil that are the hallmarks of this world. To say however that Lucius is ‘saved’ in this instance by the god Risus is stretching the point somewhat.

20. On the meaning (– ‘plot adapted from Greek’) and significance of this term cf. Mason, ‘Fabula Graecanica: Apuleius and his Greek Sources’ in Hijmans and van der Paardt (n.3 above), 1–15, at 1; Winkler (n.7 above), 184. The plays of Plautus and Terence are clearly the models that Apuleius had in mind.

21. In fact it undergoes its own metamorphosis into the religious ‘joy’ (gaudium) which permeates Book 11 and with which narrator and reader exit from the work (gaudens, ‘joyful’, is the second last word). Cf. Gwyn Griffiths, J., Apuleius of Madauros: The Isis-Book (Metamorphoses, Book XI) (Leiden 1975), 89Google Scholar and 345.

22. On the thematic significance of the priest’s speech see e.g. Penwill, J.L., ‘Slavish Pleasures and Profitless Curiosity: Fall and Redemption in Apuleius’ Metamorphoses’, Ramus 4 (1975), 49–82CrossRefGoogle Scholar, at 74f.

23. On this link between Book 3 and Book 11 cf. Nethercut, W.R., ‘Apuleius’ Literary Art: Resonance and Depth in the Metamorphoses’, CI 64 (1968), 110–19Google Scholar, at 117 n.l 1.

24. Cf. Heine (n.8 above), 31: ‘By the priest’s exhortation … [Apuleius] challenges, as it were, the reader too, … to rectify his misconception of the first ten books.’ As Lucius is saved from his asinine state, so we are saved from the error of assuming that Apuleius has merely been ‘doing a Plautus’ in his Latin version of the Greek Ass-story.

25. See Penwill (n.22 above), 50–59.

26. Cf. Fredouille, J.-C., Apulée Métamorphoses Livre XI (Paris 1975), 66Google Scholar, who notes the use of ‘vocabulaire du théâtre et du mime’ in the description of the procession at 11.8.

27. Colin, J., ‘Apulée en Thessalie: fiction ou vérité?’ Latomus 24 (1965), 330–45Google Scholar, claims that the theatre would have been the normal venue for trials in a Greek city. Whether or not this is the case, it is significant that Apuleius has the trial originally located in the forum; it is only moved to the theatre because of the pressure of numbers (3.2). We are obviously meant to notice this transition; and it is surely no accident that it reflects the classic trajectory of comedy (see Segal, E., Roman Laughter. The Comedy of Plautus [New York 1971 ], 42–69Google Scholar).

28. 11.8. Nethercut (n.23 above, 117ff.) and Hooper (n.4 above) speculate on the ways in which the personae represented in the anteludia recall Lucius’ earlier experiences. Neither offers a complete list. Nethercut concentrates on the caricature of Pegasus and Bellerophon, citing examples from the earlier books in which Lucius has compared himself and his rider to this mythical exemplar (see 6.30, 7.26, 8.16); Hooper is right about the soldier and the transvestite but wrong about the magistrate and the tame bear. The following seems to be the most plausible set of correspondences: the soldier recalls the miles of 9.39–10.1; the hunter recalls the circumstances of Tlepolemus’ death (8.4–6); the transvestite recalls the dubious sexuality of the priests of the dea Syria (addressed as puellae, ‘girls’, by Philebus at 8.26); the gladiator recalls the bandits and their blood-and-thunder lifestyle (they are actually called gladiatores by Charite at 4.24 and 26); the magistrate recalls the Festival of Laughter, in which the Hypatan magistrates begin by arresting Lucius at 3.2 and end by offering their apologies at 3.11; the philosopher recalls the philosophising ass of 10.33 and the fictive relationship established between Lucius and the philosophers Plutarch and Sextus (1.2); the bird-catcher recalls Lucius’ attempt to capture the form of a bird by magical metamorphosis (3.21–24); and the fisherman recalls Lucius’ contretemps in the fish-market (1.24f.). The tame bear dressed as a woman and the ape dressed as Ganymede represent an ironic reversal of Lucius’ state (animals disguised as humans/human in animal guise); more particularly, the bear in its cultus matronalis (‘respectable woman’s garb’) recalls the matrona pollens et opulens (‘powerful and wealthy woman’) of 10.19 whose sexual appetite proves as vast as the ass’s penis, thus pointing to the bestial sensuality that lurks beneath the veneer of respectability, while the fact that Ganymede, the archetypical male sex-object, is presented as an ape is a salutary reminder to Lucius and the reader that physical good looks and sex-appeal (in which Lucius himself was by no means deficient — see 2.5 fin.) have no intrinsic worth. Finally the ‘laughable’ (cf. rideres utrumque, ‘you would have laughed at both’, 11.8 fin.) caricature of Pegasus and Bellerophon as ass with stuck-on wings and lame old man recalls the ludicrous attempts of Lucius in Books 1–10 to give an epic flavour to his exploits by comparing them with well-known events in Greek mythology. By having the anteludia allude to Lucius’ previous experiences in this way, Apuleius establishes another nexus between Books 1–10 and Book 11: Lucius’ previous life is to his ‘rebirth’ as the anteludia are to the ploeaphesia proper — a humorous prelude, but one with a message for those ‘in the know’.

29. Similar (but not identical) is the misreading of Thiasus and his servants in Book 10. All they can see in the ass who enjoys human food and copulates with a human female is a freak whose behaviour evokes laughter as extreme as that of the Hypatans (compare 10.15f. with 3.10) and whose sexual proclivities would make a great act in the up-coming munus (10.23). Their misreading is that of those who see the ass-story as nothing but salacious comic entertainment — a Milesian tale in fact. This limited view of the work is to some extent shared by Norwood, F., ‘The Magic Pilgrimage of Apuleius’, Phoenix 10 (1956), 1–12CrossRefGoogle Scholar, at 5f., who follows Macrobius in placing this tale in the category of ‘fabulae designed solely for pleasuring the ear’ (SS 1.2.8).

30. On Apuleius’ use of scilicet in speculative assertions see Dowden, K., ‘Apuleius and the Art of Narration’, CQ 32 (1982), 419–35CrossRefGoogle Scholar, at 422ff.

31. Tatum tries to defend the devotees from this charge of error: see Tatum (n.l above), 60 and 88, and ‘The Tales in Apuleius’ Metamorphoses’, TAPA 100 (1969), 487–527Google Scholar, at 492 n.23, where the innocenna of this passage is said to refer to Lucius’ ‘credulity and inability to recognise evil’. This in itself is a strained interpretation, and does not succeed in glossing over the obvious incongruity: the Priest at 11.15 unequivocally states that Lucius ‘fell on the slippery path of youth into servile pleasures’ (lubrico uirentis aetatulae ad seruUes delapsus uoluptates), while the devotees at 11.16 assume that the divine favour shown to Lucius is evidence of the ‘guiltlessness and faith of his previous life’. See Winkler (n.7 above), 21 If., with references cited at 212 n.l2. Gwyn Griffiths (n.21 above, 257) suggests that the devotees are pronouncing a makarismos (‘blessing’) whose words ‘are a set formula’. This does not absolve them from error; rather it makes it all the clearer that their misreading is caused by too limited a perspective. James (n.7 above, 245) draws an interesting parallel with Homer’s famous ‘blameless Aegisthus’ (Od. 1.29), claiming that both Lucius and Aegisthus are ‘in possession of an innate quality by virtue of [their] nobility’. But surely this can only serve to strengthen our awareness of the gap between the bystanders’ assumption and the truth of the matter: Lucius, like Aegisthus, has not behaved as his social position requires.

32. Cf. Winkler (n. 7 above), 19: ‘[Apuleius] shows a very high consciousness … of the problems of meaning, of reading and interpreting.’

33. On witches’ use of lamps cf. 3.21 and van der Paardt (n.l7 above), 160f.

34. We will in fact learn from Fotis that she and Pamphile are ‘notorious for the practice of the black art’ (publicitus maleficae disciplinae perinfantes sumus, 3.16); Milo seems to be the only person in Hypata who can’t see the truth about his wife.

35. See further the final section below. Smith (n.7 above) rightly observes that ‘the libri which will record Lucius’ adventures are the eleven books which make up the novel as we have it’ (so also Scobie, A., More Essays on the Ancient Romance and its Heritage [Meisenheim-am-Glan 1973], 42Google Scholar; Winkler [n.7 above], 40). However Smith then goes on to claim that Diophanes is discredited by Milo’s revelations and that his prediction of gloria satis florida is proved false by Lucius being turned into a figure of fun at the Festival of Laughter and thereafter (cf. the ‘prophecy’ of the Hypatan magistrates that ‘the god [Laughter] will lovingly accompany you everywhere’ [3.11]). But the fact that Diophanes seemingly does not foresee his own troubles does not necessarily mean that he cannot accurately predict the destiny of others — cf. Fry, G., ‘Philosophie et mystique de la destinée: étude du thème de la Fortune dans les Métamorphoses d’Apulée’, QUCC 47 (1984), 137–70Google Scholar, at 149, who argues that Diophanes’ failure in his own case is punishment for revealing too much about others. That the gloria prediction is correct is evidenced by Asinius’ dream at 11.27; and the adjective florida, which we naturally take in its figurative sense in the present context, may also contain an allusion to the ‘flowery’ means whereby Lucius regains his human shape and becomes an object of veneration to the Isiac bystanders at 11.13ff. (cf. esp. 11.16, where Lucius describes himself as notus ac conspicuus, digitis hominum nutibusque notabilis, ‘noted, conspicuous, pointed out by the fingers and head-gestures of men’). Heine’s (n.8 above, 30) contemptuous dismissal of Diophanes’ prophecies as ‘faked’ and on a par with the multi-purpose oracle of the dea Syria priests at 9.8 is clearly wrong; Tatum (n.l above, 39) too is unduly negative in his treatment of them. Winkler (n.7 above, 39–44) is better but overplays the narratological complications of this scene.

36. Tatum (n.31 above, 497) draws attention to the way in which Socrates’ list of Meroe’s powers at 1.8 echoes the list of adunata in 1.3; cf. also James (n.7 above), 44. Note too how the substitution of active infinitives for the sceptic’s passives turns the later list into an emphatic refutation, as we discover in Meroe the agent who can do what the sceptic said couldn’t be done.

37. The gamesmanship involved in introducing a character called Socrates into this opening tale is well brought out by Anderson (n.2 above, 80), but the game is not merely playful. The narratological correspondence between this opening tale and Plato’s Symposium is striking (Apollodorus [Plato’s fictive narrator] said that Aristodemus said that Socrates said that Diotima said/Lucius [Apuleius’ fictive narrator] said that Aristomenes said that Socrates said that Meroe said) and helps establish the aura of philosophy perverted to which the sceptical companion objects. Note in particular how Socrates’ erotic relationship with Meroe in Apuleius parodies that between Socrates and Diotima in Plato: both Meroe and Diotima are ‘wise women’ (Meroe is saga [Met. 1.8], Diotima sophē [Smp. 201d.3]) with power over nature (for Meroe see previous note and Met. 1.9f.; Diotima kept the plague away from Athens for ten years [Smp. 201d.4f.]); Meroe’s sexual entrapment of Socrates (Met. 1.7) recalls Plato’s Socrates’ statement that Diotima ‘taught me about the things of love’ (erne ta erōtika edidaxen, Smp. 201d.5); and in both relationships the woman is the dominant partner. The anonymity of Aristomenes’ companion with whom Lucius has the introductory conversation (Met. 1.2–4) constitutes another reminiscence: Apollodorus’ companion is similarly anonymous (Smp. 173d-e).

38. Winkler’s (n.7 above, 27–33) analysis of the ‘debate’ between the sceptical companion and Lucius-as-narrator reaches a similar conclusion (33): ‘[it] applies to the credibility of the whole novel’. See also James (n.7 above), 42ff. On the character of the companion cf. Nethercut (n.23 above), 114: ‘He will never gain understanding, openness toward which alone brings the possibility of real knowledge, if also the hazard of delusion.’ However, the companion’s deficiency is not a lack of curiositas as Nethercut claims, but rather too rigid an adherence to a set of preconceived notions about quid possit oriri/quid nequeat (‘what can come about and what can’t’, Lucr. DRN 1.75f.). Closer, but not specific enough, is Scobie (n.35 above, 38): ‘These … rebukes … are obviously meant as a plea to the overcritical, dogmatic reader to maintain an open mind while reading.’ CM. Mayrhofer’s claim (‘On Two Stories in Apuleius’, Antichthon 9 [1975] 68–83CrossRefGoogle Scholar, at 75) that the sceptical companion is introduced to create a distance between author and narrator is not convincing.

39. Cf. Penwill (n.22 above), 64f.; Fry (n.35 above), 158f.

40. Examples of this kind of misreading (or a tendency towards it) are Martin, R., ‘Le sens de l’expression asinus aureus et la signification du roman apuleien’, REL 48 (1970), 332–54Google Scholar; Gwyn Griffiths, ‘Isis in the Metamorphoses of Apuleius’ in Hijmans and van der Paardt (n.3 above), 141–66; Peden, R.G., ‘The Statues in Apuleius Metamorphoses 2.4’, Phoenix 39 (1985), 380–83CrossRefGoogle Scholar; Hooper (n.4 above).

41. Compare in particular the emphasis on height (speos … hupsēlou, ‘a cave high up’/mons horridus … et in primis altus, ‘a rough mountain, exceptionally high’), dense foliage (daphnēisi katērephes, ‘shaded by laurels’, and peri d’aulē … dedmēto … makrēsin te pitussin, ‘the steading was fenced around with large pines and high-foliaged oaks’/mons … siluestribusque frondibus umbrosus, ‘a mountain shaded by forest foliage’ and conualles … nimium spinetis aggeratae, ‘valleys choked with overgrowth of thorn-bushes’), rocks (dedmēto katōrukheessi lithoisi, ‘it was fenced around with quarried stones’/saxis asperrimis et ob id inaccessis cingitur, ‘it was encircled by very rough and thus unscalable rocks’) and use as a sheepfold (entha de polla mēl’ oies te kai aiges iaueskon, ‘there many flocks, both sheep and goats, would spend the night’/caulae … ouili stabulationi commodae, ‘an opening suitable for stabling sheep’). Walsh (n. 17 above, 57f. and 58 n.l) prefers to see this description as a parody of rhetorical historiography, comparing the introduction of 4.6 (res ac tempus ipsum … descriptionem exponere flagitat, ‘the very fact that we have arrived at this point in the narrative demands that I give a description’) with Sail. BJ 17.1 (res postulare uidetur, ‘my subject-matter seems to demand’); cf. also e.g. Sail. BC 5.9, Tac. Hist. 4.5.1 and the comments of Hijmans, B.L. et al., Apuleius Madaurensis Metamorphoses Book IV 1–27: Text, Introduction and Commentary (Groningen 1977)Google Scholar, 56f. and 62. Cf. Schiesaro, A., ‘II “locus horridus” nelle Metamorfosi di Apuleio’, Maia n.s. 37 (1985), 211–23Google Scholar, at 215f., who cites Lucian’s How to Write History 20: ‘When they have nothing useful to put forward or indeed are ignorant of what needs to be said, they turn to these kinds of descriptions of localities and caves [tōn khoriōn kai antrōn ekphraseis].’

42. Even Charite’s first dream about the death of Tlepolemus (4.27: the second at 8.8 is overtly and emphatically veridical) which gives rise to the old woman’s remarks and which seems to have been proved false by Tlepolemus’ rescue eventually turns out to be correct: Tlepolemus is killed by a ‘bandit’ whose object is to steal Charite. Cf. Winkler (n.7 above), 52f.; James (n.7 above), 190.

43. Tatum (n.l above, 62–68) draws attention to what he calls the ‘remarkable dissonance’ between the narrative and its surrounds — in particular that in Book 6 between ‘the pretty allegory of Cupid and Psyche and the imminent butchery of Lucius and Charite’. One of the great values of Tatum’s study is that he carefully observes Apuleius’ book divisions and the significance of the placement of material within each book; it is especially important to do this when considering Books 4 and 6. One can therefore only regret the decision of the editors of the Groningen Commentaries on Apuleius (see Hijmans et al., n.41 above) to treat Book 4 as if it ended at 4.27.

44. See e.g. Winkler (n.7 above), 53.

45. Smith (n.7 above, 521f.) comments on the inadequacy of this description, but goes too far in castigating the narrator for ‘the shallowness and superficiality of his understanding of this material’. Bella fabella correctly describes the story as experienced by Charite and Lucius at the time; it is the knowledgeable second reader with his/her overview of the entire novel who can come back and say that the thematic importance of ‘Cupid and Psyche’ is far greater than this narrated immediate reaction would suggest.

46. For argument supporting this interpretation of ‘Cupid and Psyche’ see Penwill (n.22 above), 50–59.

47. It is instructive in this context to consider another programmatic statement which both is and is not misleading. Early on in the tale of the wicked stepmother (10.2–12), Apuleius writes: iam ergo, lector optime, scito te tragoediam, non fabulam iegere et a socco ad coturnum ascendere (‘So now, my dear reader, you must know that you are reading a tragedy, not a story, and ascending from the [comic] slipper to the [tragic] buskin’, 10.2). Commentators delight in pointing out that this story does not have a tragic ending and so does not fulfil the author’s promise. So Walsh (n.l7 above), 171: ‘This programmatic statement … is not carried through to the end of the story, which ends happily. Our author seems hardly to have known how his story was going to end when he launched it …’ This second sentence well expresses the reader’s sense of being cheated: does this author really know what he’s doing? The problem is that we have responded in the same way as Croesus did when he received the oracle about destroying a mighty empire (Hdt. 1.53.3): we have not been sufficiently circumspect in interpreting. The statement does not come right at the beginning; the ergo (‘therefore’) introduces an inference to be drawn from one of the constituents of the plot: nouerca … oculos ad priuignum adiecit. iam ergo … scito (‘The stepmother … set her eyes on her stepson. So now you must know …’). Of course we anticipate the plot of Phaedra (indeed it has been argued that Apuleius’ language in the early part of this tale intentionally recalls that of Seneca’s Phaedra — see López, V.C., ‘Tratamiento del mito en las novelle de las Metamorfosis de Apuleyo’, CFC 10 [1976], 309–73Google Scholar, at 363ff.); and so of course we feel wronged when the step-mother’s ‘plot’ is foiled by the intervention of the doctor. But this does not make the tale any less ‘tragic’. That tragedy can represent a reversal of fortune from evil to good as well as from good to evil goes back to the theory of Aristotle (see esp. Poetics 11) and the practice of Aeschylus (Oresteia) and Euripides (Alcestis, IT, Helen). When Croesus reproached the Delphic oracle for misleading him, the oracle quite rightly replied that he should have asked which empire was meant instead of jumping to conclusions (Hdt. 1.91.4). We too should be asking what kind of tragedy this tale represents instead of accusing the author of sharp practice or incompetence. It is a warning that programmatic statements in this work require careful review and revaluation; that, as with the oracle, their true meaning can only be known after the events they refer to have taken place.

48. An almost subliminal reinforcement of this notion of return is the repetition of the Geryon analogy: trigemino corpori Geryonis (‘the threefold body of Geryon’, 3.19) echoes in uicem Geryoneae caedis (‘in place of the slaughter of Geryon’, 2.32 fin.)

49. Scobie, , Apuleius Metamorphoses (Asinus Aureus) I: A Commentary (Meisenheim-am-Glan 1975), 71Google Scholar, says of mutuus nexus that it is a ‘favourite expression of Apuleius’, citing three other instances; van der Paardt (n.17 above), 137, observes that it is otherwise ‘extremely rare’. ‘Reciprocal intertwining’ is the appropriate translation for both its occurrences in the Metamorphoses (the other comes in 3.18 where it is used of Pamphile twisting the hair of the wineskins in order to perform her magic spell). Winkler (n.7 above, 188–94) devotes a good deal of discussion to the phrase but the contractual reference he imputes to it is not really appropriate here.

50. The ‘born again’ concept has so permeated Christian culture that caution is necessary when dealing with ‘rebirth’ in other religious contexts. Fredouille (n.26 above, 87) in fact denies a religious meaning here: ‘le mot [sc. renatus] n’a pas de signification mystique’. However it certainly does carry religious significance at 11.21, where the priest uses it to describe the resurrection of initiates from the uoluntaria mors (‘voluntary death’) which is part of the initiation experience. The fact that renatus here is qualified in exactly the same way as it is at 11.21 (i.e. by quodam modo, ‘in a sense’) and the fact that it is contained in a speech expressing religious fervour both suggest that Fredouille is mistaken and that we should therefore read it as 1 have indicated. Cf. Gwyn Griffiths (n.21 above), 5 If. and 258f.; Heller (n.5 above), 335.

51. Cf. Winkler (n.7 above), 213f. It is as if Christ had effected the miracle cures recorded in the Gospels by use of antibiotics.

52. On the links between Lucius’ interest in magic in Books 1–3 and his quest for Isiac salvation in Book 11 see Sandy, , ‘Serviks Vohtptates in Apuleius’ Metamorphoses’, Phoenix 28 (1974), 234–44CrossRefGoogle Scholar, at 239–42, and passages there cited. On curiositas generally see e.g. Penwill (n.22 above), 66–72; Schlam (n.6 above), passim; Sandy (n.6 above), passim; and Cooper, G., ‘Sexual and Ethical Reversal in Apuleius: The Metamorphoses as Anti-Epic’ in Deroux, C. (ed.), Studies in Latin Literature and Roman History II (Brussels 1980), 436–66Google Scholar, at 457–61.

53. Thus to say as Winkler does (n.7 above, 194) that ‘the Isiac deacon and the young Corinthian curiosus are as little alike as Lucius and his horse’ is to imply too radical a break. Through all three stages (young nobleman, ass, Isiac) Lucius is identifiably the same person as far as character is concerned.

54. On the meaning and significance of this name see e.g. Sandy (n.3 above), 136f.; van der Paardt (n. 17 above), 102.

55. Some (e.g. Walsh [n.17 above, 60f.]) see in this a link with the pose described at Apol. 33.7: interfeminium tegat et femoris obiectu et palmae uelamento (‘she conceals her genital area both by the forward thrust of her thigh and by the veil of her palm’). The fact that this is introduced by the words ad mea scripta confugit et quodam libro meo legit (‘he [sc. Tannonius Pudens] resorted to my writings and read from some book of mine’) led Todd to conclude that the book there mentioned is the Metamorphoses, which must therefore be a youthful work published prior to 158/9, the date of the Apology (Todd, F.A., Some Ancient Novels [New York 1940], 108Google Scholar). However, Apuleius is clearly not referring to the Metamorphoses since he is saying that Pudens quoted the passage ‘from my book’ in order to avoid using the word feminal; one can’t do this by reading an extract which contains the word. And in any case, the poses are hardly identical; concealment is only afforded by the thigh from a side-on view (which suggests a relief or painting rather than a statue), whereas we and Lucius are clearly being afforded a frontal view of Fotis. On the date of the Metamorphoses see Walsh op. cit. 248–51, where internal evidence is used to suggest a date after 177 AD, when Apuleius was in his fifties. For a further consideration of evidence for dating, see Gwyn Griffiths (n.21 above), 7–14.

56. Or perhaps the Aphrodite of Praxiteles. See Richter, G., A Handbook of Greek Art, 8th ed. (New York 1983)Google Scholar plates 214 and 188. The type is that of ‘Venus Bathing’. For the modern reader it is hard to eliminate the powerful image of Botticelli’s ‘Birth of Venus’, which no doubt contributed to Graves’s mistranslation (‘the Love-goddess rising from the sea’: Apuleius, , The Golden Ass, tr. Graves, R. [Harmondsworth 1950], 50)Google Scholar; cf. Amat (n.ll above), 119: ‘C’est ainsi, qu’oublieux du réalisme de la scène qu’il [sc. Lucius] dépeint, il l’évoque en un nu pictural, véritable Botticelli avant la lettre, dont la grâce rivalise avec celle de Zeuxis ou d’Apelle.’ It is interesting that the new and very welcome translation by J.A. Hanson in the Loeb Classical Library series (Apuleius, , Metamorphoses, 2 vols. [Cambridge MA and London 1989]CrossRefGoogle Scholar) similarly mistranslates: ‘Venus rising from the ocean waves’ (i.93).

57. On Lucius’ relationship with Fotis as an elegiac seruitium amoris see Sandy (n.52 above), 237f; James (n.7 above), 52f. and 68 n.34.

58. See e.g. 6.29 where lunae splendor (‘the moon’s brightness’) is introduced merely to account for the fact that the bandits see Charite trying to escape.

59. Cf. Schlam, ‘Sex and Sanctity: The Relationship of Male and Female in the Metamorphoses’ in Hijmans and van der Paardt (n.3 above), 95–105, at 104: ‘The account of Isis emphasises her womanliness, and her worshippers tend to her beauty. Lucius, as a male initiate, takes on the role of her divine consort.’

60. A link observed by Smith (n.7 above, 529) and Tatum (n. 1 above, 89–91).

61. Both are associated with roses, too. As she enters the room prior to their first bout of love-making, Fotis laeta proximat rosa serta et rosa solum in sinu tuberante (‘joyfully approached with garlands and single blooms of roses in her swelling bosom’, 2.16); while Isis of course provides the corona (‘crown’) of roses which allows Lucius to regain his human shape (11.6, 12). This ‘giving of a garland’ itself recalls 2.16, where Fotis proceeds to ‘bind me with garlands [corollis] and strew me with flowers’ — with which cf. the ‘crown [corona] intricately wrought with various kinds of flowers’ that encircles Isis’ head at 11.3.

62. These links between Fotis and Isis are somewhat different from those claimed by Sandy (n.3 above, 135f.), who argues that Fotis is represented as ‘a kind of priestess’ in a way which sets up a contrast between her and the ministers of Isis in Book 11. It is also going too far to see these links as contributing to ‘la nette prédominance de la symbolique féminine’ and ‘la féminisation de l’univers’ as claimed by Thomas (n.18 above, 194–99).

63. Cf. Penwill (n.22 above), 7If.

64. Winkler (n.7 above), who deals more fully with the ending that any other critic, reflects the prevailing imbalance of attention by devoting 24 pages to 1.1 (including 10 to the first sentence), but only 13 to 11.26–30 (of which 4 are given to the concluding sentence). The Hijmans and van der Paardt collection (see n.3 above) has two essays out of fifteen on Book 11, but one (Gwyn Griffiths, n.40 above) deals more with religious matters than literary analysis and the other (Sandy, n.3 above) pays little specific attention to the Roman closure.

65. Vesperaque quam dies insequebatur Iduum Decembrium (‘and on the evening which was to be followed by the Ides of December’, 11.26).

66. Transcurso signifero circuto Sol magnus annum compleuerat (‘The great sun, having run through the circle of the zodiac, had completed a year’, 11.26). The debate as to whether this means ‘at the end of the calendar year’, ‘one year after my initiation at Corinth’ or ‘one year after my arrival in Rome’ is documented by van der Paardt, ‘Various Aspects of Narrative Technique in Apuleius’ Metamorphoses’ in Hijmans and van der Paardt (n.3 above), 75–94, at 86 and 94 n.98, and by Gwyn Griffiths (n.21 above), 328f. The third alternative is certainly right — see Fredouille (n.26 above), 124f.

67. The only instance of nomen + cognomen in the entire novel, and a name of unquestioned authenticity. P.-W. (II.2, 1588) gives three Asinii Marcelli, all of consular rank: M. Asinius Marcellus (cos. 54 AD), Q. Asinius Marcellus (cos. suff. 96 AD) and M. Asinius Marcellus (cos. 104 AD). The name given this priest stands in striking contrast to that of the priest of Isis at Corinth, Mithras, of which Winkler (n.7 above, 245) says with some justice: ‘To give the name Mithras to the high priest of Isis … is like introducing the pope in the last chapter of a detective novel and calling him Martin Luther.’ But this is characteristic of Corinth; it is surely no more odd to have a high priest of Isis named Mithras than to have a highranking city official named Thiasus (‘Revel’)!

68. Stated not once but twice: cf. quaesticub forensis nutrito (‘nourished by a little money made in the courts’, 11.28) and stipendii forensibus fotum (‘supported by income from the courts’, 11.30).

69. This apparent change in the narrator’s birthplace has occasioned much critical comment (see below), but there is nothing in the manuscript tradition apart from some uncertainty as to spelling to suggest that Madaurensem is corrupt; see Paardt, van der, ‘The Unmasked “I”: Apuleius Met. XI27’, Mnem. 34 (1981), 96–106Google Scholar, at 98–102. Attempts such as that of Fredouille (n.26 above, 15–17) to remove the difficulty by emending the text (Fredouille proposes Corinthiensem) smack of the same kind of desperation that motivates those who try to edit the ivory gate out of Aeneid 6.898.

70. Winkler (n.7 above, 224) remarks on the fact that the narrator ends his story with the imperfect obibam (‘I was fulfilling’) rather than a preterite; this he claims ‘leaves the narrative circle enclosed: The identity of the impersonated I is never brought into contact with the present narrator of the prologue.’ But there is surely no real difficulty here; the imperfect is frequently used to indicate action commenced and continuing (cf. e.g. Kühner, R. and Stegmann, C., Ausführliche Grammatik der lateinischen Sprache, 5th ed. [Munich 1962]Google Scholar, i.l20f.), and munia … obibam is quite legitimately translatable as ‘I proceeded to fulfil the duties [and am still doing so]’. What we may have in this final word is an allusion to Ovid’s Metamorphoses; just as in the first sentence of the prologue Apuleius carefully avoids using any of the words found in Ovid Met. 1.1–4 (note figuras [‘figures’] instead of ‘formas’ [‘shapes’], in alias imagines conuersas [‘converted into other likenesses’] instead of in noua mutatas corpora [‘changed into new bodies’], uarias fabulas [‘various tales’] instead of perpetuum carmen [‘continuous song’], the emphatic first person singular of ego … conseram [‘I shall string together’] instead of Ovid’s slippery fert animus dicere [‘my spirit leads me to tell’] and the address to the reader instead of Ovid’s prayer to the gods), so at the end Ovid’s triumphant uiuam (‘I shall live’ 15.879) is replaced by the imperfect obibam — a word whose semantic range includes the meaning ‘I was dying’.

71. CD 18.18.

72. The fact that certain elements in a work of fiction parallel the known life experiences of the author does not turn the work into autobiography. The fictive narrator of Lolita, for example, emigrated from Europe to the USA at about the same time as the author (1940) and, like the author, took up an academic fellowship in literature; but one is hardly justified in deducing from this that Vladimir Nabokov himself is concealed behind the mask of Humbert Humbert. Even a work which draws as heavily on its author’s experience as A Portrait of the Artist as a Young Man is properly classed as novel rather than autobiography, fiction rather than fact: the switch from third person to first in the final chapter does not turn Stephen Dedalus into James Joyce any more than the entry of the ‘man from Madauros’ turns Lucius into Apuleius.

73. The identity of the prologue’s persona has occasioned much critical debate: see Dowden (n.30 above), 427f, and references there cited. Dowden’s own conclusion, that Apuleius has combined two kinds of comic prologues to introduce his fabula (one spoken by the playwright before he assumes his acting role, the other by a character who will take part in the play) is ingenious but unnecessarily complicating: on this interpretation the author himself speaks the first sentence, Lucius begins with exordior and the question quis ilk? (‘Who’s that?’) is addressed by the reader to the author who allows Lucius to reply on his behalf. This theory cannot really be said to overcome the problem which it is designed to address (why quis Hie? rather than quis tu?); and it is in any case far more natural to assume that the narrator speaks the first sentence too. The juxtaposition of ego tibi as the second and third words of the prologue establishes right at the outset that dialogue between the narrating ‘I’ and the listening/reading ‘you’ which will be sustained throughout the novel to the ecce (‘would you believe it’ as Graves translates) of 11.29 init. and which can only succeed if we assume that the ego is Lucius from start to finish. Winkler (n.7 above, 200–3) goes beyond Dowden to postulate a third persona distinct from both author and participating character, citing Plautus’ prologues as parallels; but this argument rests on the notion that e.g. the Mercury who speaks the prologue of Amphitruo is not the Mercury who takes part in the action of the play, which creates an artificial and ultimately meaningless distinction (it would not have the comic force it does if it weren’t spoken by this character). The narrating ‘I’ is in a position to make programmatic and interpretive statements as well as tell the story, and can switch from one to the other at will. The narration of Lolita provides countless examples.

74. Scobie (n.49 above, 74) gives ‘reminiscent of the market-place’ for forensis, in which he is followed by Winkler (n.7 above, 181) who punningly translates ‘bazaar language’. Elsewhere in the Met., however, the word is associated with the legal rather than the commercial aspect of the forum (9.6, 10.33, 11.28, 11.30); any uncertainty a first-time reader may have as to its meaning here should be resolved by the time s/he reaches the end of Book 11.

75. That Lucius already has command of the Latin language is indicated inter alia by the fact that he understands what the miles is saying at 9.39 while the market-gardener doesn’t, by his record of Apollo’s conspicuous use of Latin in the old woman’s narration of ‘Cupid and Psyche’ (4.32 fin.), and by the Roman juridical flavour of the speeches at the mock trial in Hypata, where the word Quirites (employed by both prosecutor and defendant to address the jury) echoes the prologue and suggests that its user is now fully conversant with studiorum Quiritium indigena sermo (‘the native language of Roman studies’).

76. For other, less sustained, examples see Summers, R.G., ‘Justice and Apuleius’ Metamorphoses’, TAPA 101 (1970), 511–31Google Scholar, at 515.

77. On the levels of meaning in this oracle see Penwill (n.22 above), 78 n.28. It is yet another instance of a programmatic statement (gods share with narrators the superior knowledge which enables them to make such statements) which both does and does not mislead — cf. n.47 above.

78. Smith (n.7 above, 532) correctly observes that it is not the narrator who identifies Lucius as the ‘poor man from Madauros’ but Osiris, and goes on to contend that the identification ‘may have a secret meaning unknown to anyone in the story’. In the sense that it constitutes another model for misreading, that is true; it is part of the dialogue between author and reader in which Lucius as fictive narrator can have no role. As in the case of Aristomenes’ companion’s scepticism, Milo’s denigration of Diophanes, the old woman’s description of ‘Cupid and Psyche’ and his own response to it at the time of its telling, and the reaction of the Isiac devotees to his retransformation, Lucius merely records; it is Apuleius who is asking us to interpret.

79. The narratological problems raised by identifying Apuleius with Lucius on the basis of Osiris’ oracle are illustrated by Scobie’s complaint (Aspects of the Ancient Romance and its Heritage [Meisenheim-am-Glan 1969], 34f.): ‘His [Lucius’] fictional identity is shattered by Apuleius’ own identification with him at 11.27 … [This] not only creates retrospective doubt in the reader’s mind about the fictional authenticity of his hero-narrator, but the structural logic of the narrative remains uncoordinated thanks to a fluctuating combination of fact and fiction.’ Cf. Dietrich, B.C., ‘The Golden Art of Apuleius’, G&R n.s. 13 (1966), 189-206Google Scholar, at 203: ‘Lucius, the hero, plays two cunning roles, one autobiographical and the other fictional, so ingeniously fused and interchanged that the bewildered reader can no longer separate truth from invention.’ Puzzlement and/or accusations of incompetence are not uncommon among those who accept the identification: see e.g. Perry, B.E., The Ancient Romances: A Literary-Historical Account of their Origins (Berkeley and Los Angeles 1967), 242Google Scholar and 250; Walsh (n.17 above), 184 and 187; Gwyn Griffiths (n.21 above), 6. But if we find the text bewildering, it should suggest that something is wrong with the way we are reading it.

80. As indeed is one which tries to make historia and fabula mean much the same thing: Scobie (n.35 above), 42.

81. There is thus some value in MacKay’s description of the work as ‘an extended myth’, although to claim that ‘the moral of the story … is that power, wealth and all self-seeking are fatal to the soul’ is to over-simplify. See MacKay, L.A., ‘The Sin of the Golden Ass’, Arion 4 (1965), 474–80Google Scholar, at 479 and 477.

82. See Ebel, H., ‘Apuleius and the Present Time’, Arethusa 3 (1970), 155–76Google Scholar, who speaks of ‘the universal fact (of which “magic” is as much an expression as a cause) that all appearances are unreliable and all “realities” transient’ and of ‘the unexpected twists and turns of a plot whose movement exemplifies the nature of what it portrays’ (163). In a world in which we can never step into the same river twice appearances are deceptive (cf. n.18 above) and human perceptions consequently fallible. So the ass complains in 7.2f. that in a world dominated by Fortuna, reputation frequently runs counter to truth; and at 10.33 we find the same animal fulminating against the corrupt and erroneous judgements passed on innocent individuals. The conclusion of this latter passage, in which an imaginary reader is represented as interjecting ecce nunc patiemur philosophantem nobis asinum (‘so now we’re going to let an ass give us a philosophy lecture, are we?’), has its own irony in that once we have finished Book 11 we realise that that is exactly what we have been doing.